Learning When to Speak: Latency and Quality Trade-offs for Simultaneous Speech-to-Speech Translation with Offline Models.
Liam DuganAnshul WadhawanKyle SpenceChris Callison-BurchMorgan McGuireVictor B. ZordanPublished in: INTERSPEECH (2023)
Keyphrases
- language acquisition
- trade off
- accurate models
- learning systems
- speech recognition
- probabilistic model
- prior knowledge
- learning algorithm
- speech signal
- learned models
- learning process
- reinforcement learning
- high quality
- real time
- hidden variables
- structured prediction
- computational models
- speech synthesis
- text to speech
- grapheme to phoneme conversion
- learning rules
- audio visual
- neural network